From: Martin Stromberg <Martin DOT Stromberg AT lu DOT erisoft DOT se>
Message-Id: <200009151122.NAA29116@lws256.lu.erisoft.se>
Subject: Re: Bug 000323
To: djgpp-workers AT delorie DOT com
Date: Fri, 15 Sep 2000 13:22:31 +0200 (MET DST)
In-Reply-To: <Pine.OSF.4.20.0009150929560.3790-100000@sirppi.helsinki.fi> from "Esa A E Peuha" at Sep 15, 2000 10:45:02 AM
X-Mailer: ELM [version 2.5 PL3]
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Reply-To: djgpp-workers AT delorie DOT com
Errors-To: nobody AT delorie DOT com
X-Mailing-List: djgpp-workers AT delorie DOT com
X-Unsubscribes-To: listserv AT delorie DOT com
Precedence: bulk

> I think it would be good to pass the append bit to the fsext mechanism.
> 
> Of course, open(filename, O_APPEND) should work like
> fopen(filename, "a"), so if you feel like implementing that, it would
> be nice; that's not so easy to do, though.

Ah. Is there a document somewhere that explains open() and O_APPEND
and the interactions with lseek() and company?

> Why would you want not to have the file pointer at the end of the
> file?  But if you pass O_APPEND to open (to let fsext know about the
> append bit), it will seek to end anyway, so this really is redundant.

Why would I? It'll be fseek()ed when you try to fwrite() or fprintf()
to the stream...


Right,

						MartinS